Introduction

The sub-$200 market-segment has been NVIDIA's problem area in terms of price-performance competitiveness with AMD, even though the company has had higher volumes. Most games released for the competitive gaming crowd run great on sub-$300 graphics cards, and it's only blockbuster AAA titles with cutting-edge production designs that prompt people to invest in faster graphics solutions, where NVIDIA has established an unbeatable lead.

NVIDIA is getting hawkish and wants itself a bigger slice of the sub-$200 market-segment targeting e-Sports players. The company launched the $139.99 GeForce GTX 1050 Ti and $109.99 GTX 1050 earlier this month. The two SKUs are based on NVIDIA's smallest implementation of its "Pascal" GPU architecture, the GP107 silicon. This tiny chip packs up to 768 CUDA cores, 48 TMUs, 32 ROPs, and a 128-bit GDDR5 memory interface, holding 4 GB of memory on the GTX 1050 Ti.

These are still "Pascal" CUDA cores that tick at 1.35-1.45 GHz. NVIDIA's decision to go with a 128-bit wide memory bus shows that the GTX 1050 Ti has been built to a cost (with no more than four memory chips), which prepares NVIDIA for a price war with better-endowed, but costlier to make AMD offerings. NVIDIA managed to get the power consumption of the GTX 1050 Ti below the 75W mark, which makes it capable of sustaining itself on slot power alone. Its nearest rival from AMD, the Radeon RX 470 4GB, needs an additional 6-pin PCIe power connector to feed its 120W TDP setup.



In this review, we are testing the Gigabyte GeForce GTX 1050 Ti G1 Gaming, a premium custom-design graphics card by Gigabyte that combines a factory-overclocked GTX 1050 Ti implementation with the company's signature dual-fan cooling solution, which turns its fans off when the GPU is idling, and a custom-design PCB that features an additional 6-pin PCIe power connector to help bolster the card's overclocking headroom.